hey c! curious if you use AI for research and figuring out new sites and such if not for writing. do you?
Hello! I have a couple of things to say —
My blanket rule is that there should be absolutely no generative AI use for writing. I would like a spellcheck and grammar thing, and grammarly used to be good before it broke.
Personally, I’ve not used it for research like to direct me to websites all that much. I did use it while I was working to compile easy lists of where I could find raw data, but since my work has stopped being that, that use has pretty much petered out. I’ve not found it to be very useful for archival research.
At the same time I also think we need to be more precise with what we mean when we say “using AI”. This is not directed at you, anon, but more generally something I’ve been thinking about, because using genAI and language models is very very different from using automation or machine learning for things like pattern recognition. And using AI catch-all term makes the conversation harder especially when we want to talk about ethical use and regulation.
Lots of archives I know have been using automation or things like text recognition to make their search engines better, which has been so helpful in finding documents. One of the collections I regularly use also added a whiteboard sort of thing some months ago where you can get an AI assistant to convert a manuscript to typed text that you can then edit. Another archive also did this along with allowing crowdsourced annotations. In these cases it’s also been really exciting to see what collections might do, and how they navigate privacy and data protection concerns.
The TLDR version is that no, I don’t directly use it because I’ve not found it to be very good at doing what I need it to. But I’m definitely indirectly using for more embedded things that are like smaller steps in the research process, because it’s increasingly part of the systems I do normally use.
